func MakeData(start float64, end float64, inc float64) []float64 {
var data []float64
rng.EnvSetup()
T := rng.DefaultRngType()
r := rng.RngAlloc(T)
for x := start; x < end; x += inc {
y0 := math.Exp(x)
sigma := inc * y0
dy := randist.Gaussian(r, sigma)
data = append(data, []float64{x, y0 + dy, sigma}...)
}
return data
}
func TestBspline(t *testing.T) {
var n int = 200
var ncoeffs int = 12
var nbreak int = ncoeffs - 2
rng.EnvSetup()
r := rng.RngAlloc(rng.DefaultRngType())
// allocate a cubic bspline workspace (k = 4)
bw := bspline.Alloc(4, nbreak)
B := vector.VectorAlloc(ncoeffs)
x := vector.VectorAlloc(n)
y := vector.VectorAlloc(n)
X := matrix.MatrixAlloc(n, ncoeffs)
c := vector.VectorAlloc(ncoeffs)
w := vector.VectorAlloc(n)
cov := matrix.MatrixAlloc(ncoeffs, ncoeffs)
mw := multifit.LinearAlloc(n, ncoeffs)
fmt.Printf("#m=0,S=0\n")
// this is the data to be fitted
for i := 0; i < n; i++ {
xi := (15.0 / (float64(n) - 1)) * float64(i)
yi := math.Cos(xi) * math.Exp(-0.1*xi)
sigma := 0.1 * yi
dy := randist.Gaussian(r, sigma)
vector.Set(x, i, xi)
vector.Set(y, i, yi+dy)
vector.Set(w, i, 1.0/(sigma*sigma))
fmt.Printf("%f %f\n", xi, yi+dy)
}
// use uniform breakpoints on [0, 15]
bspline.KnotsUniform(0.0, 15.0, bw)
// construct the fit matrix X
for i := 0; i < n; i++ {
xi := vector.Get(x, i)
// compute B_j(xi) for all j
bspline.Eval(xi, B, bw)
// fill in row i of X
for j := 0; j < ncoeffs; j++ {
matrix.Set(X, i, j, vector.Get(B, j))
}
}
// do the fit
_, chisq := multifit.Wlinear(X, w, y, c, cov, mw)
dof := float64(n - ncoeffs)
tss := stats.Wtss(w.Data_(), w.Stride(), y.Data_(), y.Stride(), n)
rsq := 1.0 - chisq/tss
fmt.Printf("chisq/dof = %e, Rsq = %f\n", chisq/dof, rsq)
fmt.Printf("#m=1,S=0\n")
for xi := 0.0; xi < 15.0; xi += 0.1 {
bspline.Eval(xi, B, bw)
_, yi, _ := multifit.LinearEst(B, c, cov)
fmt.Printf("%f %f\n", xi, yi)
}
}
